Researches On OSGI-based Distributed Service Sharing Middleware

The sensing layer in Internet of Things(IoT) contains a large number of heterogeneousembedded sensing devices, which is the foundation of the IoT system. Distributed embeddeddevices can make accurate perception of the objective world and efficient control. Asinformation exchanging link between underlying sensing devices and IoT application system,the sensing layer middleware provides unified interfaces to deploy, manage and invokesensing services. Ensure the sensing layer network provide a wide range of dynamic andscalable sensing services, a fast and convenient service development and management mode,has become one of the key point that needs to be solved in the sensing layer middleware.Based on OSGi framework, a service oriented middleware for IoT is proposes, whichcalled OSGi based distributed service sharing middleware. This middleware platform makesuse of the OSGi service oriented advantages, dynamically publishing and banding sensingservices which are provided by sensing layer. Change the dependence of IoT application layeron the sensing layer, and make IoT application only acting with services not with specificsensing devices. This breaks the dependence between IoT application layer and the sensinglayer. IoT applications can design and build in dynamic, modular way by taking advantage ofservice-oriented design model.DssOSGi first focus on services sharing extension for distributed OSGi. DssOSGi can letembedded sensing device get access to other services provided by networked embeddeddevices easily. Thus sensing devices could complete the related business processes orco-operating. Second, DssOSGi enables transparent real-time event delivery and notice,which making the sensing devices quickly respond to the upper application requests.This paper proposed a kind of dynamic service composition algorithm DscGOM basedon distributed OSGi. The algorithm classifies and adaptively matches the existing services,dynamically generates combined service to deal with different application demands.Meanwhile, due to the impact of mobility and battery life, the embedded sensing deviceswould be instability when binding and invoking services provided by these devices. Toovercome this problem, DscGOM establish a composite service redirection mechanism,which ensures that we could quickly recover from a failure service during the execution of theservice sequence.Finally, we carry out experiments on DssOSGi middleware platform and simulateDscGOM service composition algorithm. Performances are tested and evaluated in terms of the response time, resource consumption.